Cellular and molecular mechanisms involved in T cell-mediated autoimmunity against immunologically privileged retinal antigens are being studied. The questions are aimed at elucidating the natural development and maintenance of self-tolerance to retinal antigens, and at defining the processes that lead to their pathological breakdown. The goal is to use this knowledge for designing novel and rational strategies for immunotherapy. The experimental approaches utilize the model of experimental autoimmune uveoretinitis (EAU), which resembles immune-mediated uveitic diseases in humans that can lead to blindness. EAU is induced in mice and rats by immunization with retinal antigens such as IRBP, Arrestin (S-Ag), or their component peptide epitopes, or by infusion of cultured lymphocytes that recognize these antigens. The mechanisms controlling disease susceptibility and pathogenesis are being defined at the genetic, developmental, and immunological levels. Novel approaches to disease regulation are devised based on these findings. Natural tolerance to IRBP is being investigated by using IRBP deficient mice. We found that thymic selection plays a major role in tuning the threshold of susceptibility to ocular autoimmunity and that the uveitis-susceptible individual in fact displays a demonstrable level of tolerance to the target antigen involved in disease. The mechanism involves elimination of highly IRBP reactive T cells and release of CD4+CD25+ regulatory T cells by the thymus. Furthermore, CD4+CD25+ regulatory cells may explain resistance to EAU of some poorly susceptible strains, as depletion of CD4+CD25+ cells in several such strains permits EAU induction. These results suggested that thymic tolerance has a greater role than hitherto realized in setting the threshold of immune and autoimmune responsiveness to sequestered retinal antigens. In contrast, we believe - based on data generated in IRBP?transgenic mice - that peripheral tolerance is an underutilized mechanism that may be exploited therapeutically. Based on this conclusion, we are studying therapeutic induction of tolerance in adult mice by hydrodynamic IV vaccination with a naked DNA plasmid encoding an IRBP fragment. This method requires only a single vaccination with 10 ?g of DNA. It is highly effective in preventing disease, but only moderately effective in reversing disease elicited by already primed T cells. Protection lasts at least 10 weeks after vaccination, appears not to require apoptosis by the major pathways (Bcl-2 or Fas mediated) or IL-2-reversible anergy, but may involve regulatory cells, as post-vaccination depletion of CD25+ cells reverses protection. Notably, vaccination with high doses of DNA, while protective from actively induced disease, may by itself elicit an IFN-g response and induce a mild, transient ocular inflammation, suggesting that this approach needs to be carefully optimized. Another approach to induce peripheral tolerance is use of altered peptide ligands (APL) of the uveitogenic epitope that bind the MHC and/or TCR with reduced avidity. Two candidate APLs, that do not induce EAU when used in what would constitute a uveitogenic immunzation regimen, but protect from EAU induced by the native epitope, were identified and are being studied. Innate immune responses to microbes, that create a pro-inflammatory environment, are thought to play a role as environmental triggers of autoimmunity. For this reason we have made efforts to delineate the role of adjuvant effects in EAU. Pertussis toxin (PT) enhances autoimmune disease in experimental models. PT administered at the time of immunization enhances the Th1 response, which underlies EAU pathology. In contrast, cholera toxin (CT), which shares structural, but not functional, homology with PT, appears to inhibit EAU by altering the cytokine profile of the response to IRBP. Rather than regulatory cells, the mechanism appears to involve channeling of the effector response to a nonpathogenic phenotype. Innate IL-4 production is detected within hours after treatment and is required for the protective effect. As another approach to innate immunity as a contributing factor to EAU, we are studying the dependence of EAU induction on signaling through the MyD88/TOLL receptor pathway. Use of MyD88, TLR2, TLR 4, TLR 9 as well as IL-18 and IL-1RI gene knockout mice revealed that although MyD88 pathway is necessary for EAU to be induced, TLR2, TLR 4, TLR 9 signaling is either unnecessary or redundant, as all 3 TLR knockouts were fully susceptible, as were IL-18 KOs. In contrast, IL-1RI deficiency conferred resistance, indicating that requirement for IL-1 signaling is necessary and nonredundant for EAU induction. Dendritic cells (DC) are central antigen presenting cells (APC) in induction of immune and autoimmune responses. We are developing an EAU model where disease is induced with in vitro-matured, antigen pulsed DC. We are able to obtain large numbers of splenic DC from mice injected hydrodynamically with Flt-3L and purify them to homogeneity. When matured (by culture with LPS and anti-CD40), pulsed with a uveitogenic IRBP peptide and injected into mice, these DC induce vigorous immune responses and elicit typical EAU-like ocular inflammation in mice. This system is an alternative to active immunization with uveitogenic proteins in complete Freund?s adjuvant and will permit to study various manipulations and agents for their ability to regulate activated DC capable of inducing EAU. In collaboration with the Oppenheim group at NCI we demonstrated that lymphocytes and immature dendritic cells exhibit chemotactic responses to the retinal antigens S-Ag and IRBP. The chemokine receptors CXCR5 and CXCR3 mediated the chemotactic effect of IRBP, while only CXCR3 was required for the S-Ag signal. We hypothesize that these responses may underlie attraction of immunocytes to sites of tissue damage and may thus facilitate processes of tissue repair and regeneration. When natural control mechanisms fail, or are subverted, such responses may lead to development of an adaptive response of the Th1 phenotype, leading to autoimmune tissue pathology. EAU, similarly to uveitis in humans, is genetically controlled by MHC as well as non-MHC genes. We have developed a ?humanized? model of EAU using HLA-transgenic mice in which murine MHC genes have been replaced by human MHC genes. We are studying the epitopes recognized by HLA Tg mice as an approach to identifying antigenic regions of the molecule involved in human uveitis in single HLA transgenic and double HLA-transgenic mice, that combine a susceptible HLA DR and HLA DQ molecule. Interestingly, HLA DR3 Tg mice develop strong EAU when immunized with S-Ag, which induces little or no uveitis in WT mice, but elicits responses in humans, and recognize the same region of S-Ag that is recognized by uveitis patients. Double-Tg mice develop faster and stronger disease than their single-Tg parents. And respond to additional antigenic regions, not recognized by either of the single HLA-Tg parental strains. This indicates that the HLA molecules interact in determining susceptibility. These results validate EAU as a model of human uveitis and promise to yield new information on epitopes that might be uveitogenic in humans.